决策树分类法及其在土地覆盖分类中的应用

基于决策树分类算法在遥感影像分类方面的深厚潜力,探讨了3种不同的决策树算法(UDT、MDT和HDT)。首先对决策树算法结构、算法理论进行了阐述,然后利用决策树算法进行遥感土地覆盖分类实验,并把获得的结果与传统统计分类法进行比较。研究表明,决策树分类法有诸多优势,如:相对简单、明确、分类结构直观,另外,与以假定数据源呈一固定概率分布,然后在此基础上进行参数估计的常规分类方法相比,决策树属于严格“非参”,对于输入数据空间特征和分类标识具有更好的弹性和鲁棒性(Robust)。

The Decision Tree Classification and Its Application Research in Land Cover

Decision tree classification algorithms have significant potential for remote sensing data classification. In this paper, three different types decision tree classification (UDT, MDT and HDT)are presented. First, the paper discussed the algorithms structure and the algorithms theory of decision tree. Second, decision tree algorithms were used to make land cover classification from remotely sensed data, and the results were compared with conventional statistics classification. The results of this research showed that decision trees have several advantages for remote sensing applications by virtue of their relatively simple, explicit, and intuitive classification structure. In addition, decision tree algorithms are strictly nonparametric and, therefore, without assumptions regarding the distribution of input data the methods are flexible and robust with respect to general classifications among input features and class labels.